How can I convert from jitter() amounts to position_jitter() widths in ggplot2?

I'm updating a package that uses base graphics to one that uses ggplot2 graphics. In the base graphics version, users could supply a value for jj, which would then get scaled and passed to a jitter() function. The code to jitter x values looked like this:

degree.of.jitter <- (jj/200) * diff(x.values.range)
jitter(x.values, amoun开发者_Go百科t = degree.of.jitter)

I'd like the ggplot2 version to achieve the same visual jittering effect as the base graphics version. But, I'm not sure how to re-scale my existing degree.of.jitter so I can pass it directly to position_jitter() and achieve the same visual results:

position_jitter(width = MysteriousScalingFunctionOfCompleteMysteryWhoseInnerWorkingsIHaveYetToSpecify(jj))

Does anyone have any suggestions for what the appropriate scaling of jj would need to be?

---

Under the hood, width is passed as the amount argument to jitter, so you should be able to just use

position_jitter(width = degree.of.jitter)

where degree.of.jitter is defined the same way as before.